Send Orders of Reprints at reprints@benthamscience.net
18
The Open Mechanical Engineering Journal, 2013, 7, 18-26
Open Access
Longitudinal Residual Stress Analysis in AA2024-T3 Friction Stir Welding
Pierpaolo Carlone* and Gaetano S. Palazzo
Department of Industrial Engineering, University of Salerno, Via Ponte Don Melillo 1, 84084, Fisciano (SA), Italy
Abstract: Friction Stir Welding (FSW) is an innovative solid-state joining process, which is gaining a great deal of
attention in several applicative sectors. The opportune definition of process parameters, i.e. minimizing residual stresses,
is crucial to improve joint reliability in terms of static and dynamic performance. Longitudinal residual stresses, induced
by FSW in AA2024-T3 butt joints, have been inferred by means of a recently developed technique, namely the contour
method. Two approaches to stress measurement have been adopted; the former is based on the assumption of uniform
material properties, the latter takes into account microstructural effects and material properties variations in the welding
zones. The influence of process parameters, namely rotating and welding speeds, on stress distribution is also discussed.
Keywords: AA2024-T3, friction stir welding, contour method, residual stress, material properties.
1. INTRODUCTION
Medium to high strength aluminum alloys, such as 2xxx,
6xxx, and 7xxx series, are currently considered of great
interest in the transport industries. In particular, the
precipitation hardenable AA2024 (Al-Cu) alloy is gaining
considerable attention for the realization of barrier beams or
fuselage panels. In this context, remarkable research effort is
focused on the application of the Friction Stir Welding
(FSW) process, as a suitable alternative to fusion welding
processes. Indeed, the poor dendritic microstructure and the
high porosity in the weld zone, induced using conventional
techniques, strongly reduce the mechanical behavior of the
assembly. Furthermore, the reduction of production costs
and weight and the increase of strength and damage
tolerance with respect to riveted lap joints make FSW a very
attractive process to the aerospace industry.
FSW is a solid-state welding process, developed and
patented by The Welding Institute (TWI) of Cambridge in
1991. Following the early successful applications in
aluminum welding, FSW has been applied to other
engineering materials, such as copper, steel, titanium, and
metal matrix composites. During the process a nonconsumable rotating tool, constituted by a shoulder and a
pin, is plunged between the adjoining edges of the parts to be
welded and moved along the desired weld line. The
combined rotation and translation of the tool locally increase
the work piece temperature, due to heat generated by
frictional effects and plastic deformation. The induced
softening allows the processing material to flow around the
pin, from the front (leading edge) to the rear (trailing edge)
according to complex patterns, resulting in a solid state weld
[1]. Temperature increase and high strain rate deformation
lead to the formation of micro-structurally different zones:
the nugget or stir zone (NZ) in the center of the weld,
*Address correspondence to this author at the Department of Industrial
Engineering, University of Salerno, Via Ponte Don Melillo 1, 84084,
Fisciano (SA), Italy; Tel: 0039 089 964320; Fax: 0039 089 964307;
E-mail: pcarlone@unisa.it
1874-155X /13
surrounded by the thermo-mechanical affected zone (TMAZ)
and the heat affected zone (HAZ). Nowadays, a deeper
understanding of static strength as well as of fatigue
behavior of FSWed assemblies is highly desired for a wider
implementation of the technique in safety-critical
components. It is commonly accepted that the
aforementioned properties mainly rely on residual stresses,
microstructure, and microhardness. Even if FSW residual
stresses are generally lower if compared to conventional
welding processes, an accurate knowledge of their
distributions is crucial to numerically investigate buckling
behavior [3-5] as well as crack growth and fatigue response
[6-13] of welded assemblies. In this sense some results have
already been presented in the inherent literature, regarding,
for instance, AA5083 [14-16], AA6082 [6,17,18], AA2195
[13], AA2219 [19], and AA2024 [8-12,17,18,20,21] FSW. In
more details, Peel et al. investigated the influence of the
welding speed on microstructural, microhardness and
residual stress on 3 mm thick AA5083 FSWed butt joints.
The stress distribution was inferred by means of synchrotron
X-rays, evidencing a tensile status in both the longitudinal
and transverse direction. Moreover an increase of the tensile
peak was found increasing the welding speed [14]. FSW
process was simulated using a fully thermo-mechanical
model in [15], declaring good agreement with available
results [14]. The influence of dynamical recrystallization
effects on residual stress distribution was also suggested.
Synchrotron X-ray radiation was employed by Lombard et
al. in [16] to analyze 6 mm thick FSWed plate. The typical
residual stresses M-shaped profile was detected, confirming
also the dominant role of the welding speed on the tensile
peak. Three different aluminum alloys (5754-H111, 2024T3, and 6082-T6) were joined using FSW in [17]. Residual
stress measurements were carried on using the slitting (crack
compliance) method. Two different inverse methods, namely
the linear elastic fracture mechanics (LEFM) and finite
element (FE) methods, were implemented, providing similar
results. Reported data highlighted the correlation between
the stress profile and the microstructural changes in the
processing material. The hole drilling technique was applied
2013 Bentham Open
Longitudinal Residual Stress Analysis in AA2024-T3 Friction Stir Welding
in [18] to measure through the thickness residual stress in
four aluminum alloys, namely AA1050-O, AA2024-T4,
AA6082-T6, and AA7075-T6, welded according to three
different processing conditions. Compressive stresses were
found on joint surface in correspondence of the advancing
side, replaced by tensile stress inside the joint. An
experimental residual stress analysis on 12 mm thick
AA2219-T6 plates, butt joined by FSW, was carried on by
Xu et al. using the hole-drilling strain-gauge method [19].
Different stress profiles were detected at the top and bottom
surfaces, following, respectively, a M-shape and a V-shape.
Moreover, an increase of the tensile peak was highlighted in
defected joints.
As far as AA2024 FSW is regarded, several studies have
been focused on the analysis of residual stresses, using
numerical and experimental approaches. Two process
settings, defined as hot and cold weld, were used by Deplus
et al. to weld 2 mm thick AA2024-T3 plates in the butt joint
configuration [17]. Longitudinal stresses were measured
using the slitting method, evidencing a more severe residual
status assuming cold weld parameters. The adopted method,
however, did not allow for the detection of stress variations
through the thickness. The same technique was used by
Milan et al. to investigate longitudinal and transverse
residual stresses in 3.2 mm thick AA2024-T3 plates [20]. A
higher tensile residual stress was measured on the advancing
side of the weld and related to the larger heat input, resulting
from the higher relative speed between the tool and the
material. Sutton et al. measured longitudinal, transverse and
normal stresses distributions in 7 mm thick AA2024 plates in
the T3 condition [21], by means of neutron diffraction (ND).
The longitudinal stress was found to be the largest tensile
component. The analysis was performed assuming one
processing condition; the influence of process parameters
was not investigated. Despite the relevant data and intriguing
conclusions reported in the cited papers, it should be noted
that most of the cited works explored very few, if not unique,
combinations of process parameters. Moreover, residual
stresses were measured in selected locations of the joined
plates or considering average values due to intrinsic limits of
the employed techniques.
In this paper results provided by an experimental
investigation on longitudinal residual stresses in 4 mm thick
AA2024-T3 FSWed butt joints are reported and discussed.
The contour method has been applied to map the stresses
acting on a section normal to the weld line. The aim of the
work is to investigate the influence of process parameters,
namely rotating and welding speed, on stress distribution,
considering also the influence of microstructural changes on
the computed stress field. Two stress maps have been
inferred for each test, assuming, respectively, constant
material properties and properties variations in the different
welding zones (NZ, TMAZ, HAZ). The extension of each
zone has been defined by means of a numerical-experimental
approach, following experimentally calibrated criteria based
on temperature, dynamic viscosity, and grain size values.
Local material properties have been quantified using
ultrasonic analysis and then employed for stress calculation.
The paper is structured as follows: in Section 2 details
regarding material, manufacturing process, and analysis
procedures are reported, while in Section 3 results are
The Open Mechanical Engineering Journal, 2013, Volume 7
19
exposed and discussed. Finally, Section 4 deals with relevant
findings of the work.
2. MATERIALS AND METHODS
2.1. Material and Welding Process
In the present investigation, AA2024-T3 aluminum rolled
plates (100 mm x 30 mm x 4 mm) were joined by FSW. The
nominal composition (wt%) of the base material is as
follows: 3.8-4.9 Cu, 1.2-1.8 Mg, 0.3-0.9 Mn, 0.5 Si, 0.5 Fe,
0.25 Zn, 0.15 Ti, 0.1 Cr, balance Al. The considered material
has been subjected to a solution heat treatment, followed by
cold working and natural ageing. Welds have been executed
normally to the rolling direction on a MCX 600 ECO
machining center. The process set up is shown in Fig. (1).
Fig. (1). Friction stir welding process set up (A: plates in reference
pocket, B: tool and tool-holder, C: Inclined base plane imposing the
tilt angle, D: clamping system).
A full factorial design of experiments has been adopted,
assuming respectively five and three levels for the rotating
speed (800, 1000, 1200, 1400, 1600 rpm) and feed rate (35, 70, 140 mm/min). Tilt angle and pin penetration have
been defined as 2° and 0.2 mm, following literature
indications. An AISI1040 quenched steel tool (56 HRC)
consisted of a 20 mm diameter shoulder with a conical
unthreaded pin has been employed. Pin dimensions are:
height 3.80 mm, major diameter 6.20 mm, and cone angle
30°. Preliminary visual inspections and microscopic
observations of the joint section showed the presence of
internal defects, such as tunnel defect and kissing bonds at
rotating speeds lower than 1400 rpm, significantly reducing
the process window. Taking into account that defect analysis
is not the main focus of the present investigation, detailed
information are not herein reported. Omitted data, analysis,
and related discussions can be found in [22].
2.2. Welding Zones Identification and Characterization
In this paragraph, numerical and experimental methods
used to individuate the different welding zones (NZ, TMAZ,
HAZ) and to evaluate the related properties are reported. As
far as AA2024-T3 FSW is concerned, it is relevant to notice
that extension, microstructure and mechanical properties of
each zone are significantly affected by dynamic
recrystallization, dissolution, coarsening, and re-precipitation
phenomena, whose kinetics are strongly influenced by
20 The Open Mechanical Engineering Journal, 2013, Volume 7
thermal effects coupled to plastic deformations [23].
Conventional optical metallographic procedures allows for a
clear detection of NZ as well as of TMAZ. Indeed, the
former is interested by continuous dynamic recrystallization
phenomena, resulting in an evident grain refinement with
respect to the parent material. In the latter a variable amount
of grain distortion, without significant size changes, is
generally observed [2,10,20,22]. HAZ boundaries are quite
difficult to observe using the aforementioned techniques,
since no relevant modifications of grain size and shape are
induced. However, the temperature field experienced by the
processing material during the welding strongly affects
precipitates distribution and properties, influencing the
microhardness (HV) distribution. As a consequence, the
peculiar HV trend in the HAZ, allows one to indirectly
evaluate the HAZ extension [23]. Experimental data have
been used to calibrate some criteria, based on temperature,
dynamic viscosity, and grain size considerations,
implemented in a CFD process model, presented and
validated in [22,24], providing an automatic definition of
welding zones boundaries.
Taking into account what aforementioned, NZ
boundaries have been defined, for each processing
conditions, in correspondence of the (numerically computed)
onset of recrystallization and grain size change [24]. A
criterion based on dynamic viscosity calculation, as reported
in [25,26], has been adopted to define TMAZ extension,
assuming a reference value of 4E6 Pa·s in correspondence of
the separation between TMAZ and HAZ [22]. In Fig. (2) a
comparison between the numerically computed and
experimentally observed NZ and TMAZ has been reported,
showing also a contour of the grain size distribution in the
NZ. An indirect definition of HAZ - BM boundaries has
been adopted, considering, as generally accepted, that no
relevant phenomena take place in the considered material
below a threshold temperature [23], herein defined as 250°C
considering microhardness results presented in [22].
Regarding material characterization, it is relevant to
emphasize that the method employed for residual stress
evaluation requires elastic properties only. The local elastic
Carlone and Palazzo
modulus has been evaluated by ultrasonic time of flight
(TOF) measurements. The used device was a
PANAMETRICS 5058PR high voltage pulser-receiver,
working in a pulse-echo mode. A V544 ultrasonic probe,
generating 10 MHz longitudinal waves, has been used. Other
parameters have been defined by trial and error as follows:
100 damping, 200 V pulse height, 30 dB attenuation, and
0.1 MHz HP filter. Time of flight measurements have been
performed in correspondence of the NZ, TMAZ and HAZ on
both sides of the weld. The local elastic modulus has been
then inferred according to the following equation:
Vl =
E(1 v)
(1+ v)(1 2v)
(1)
being Vl the ultrasonic longitudinal speed, the density and
the Poisson ratio. Three measurements have been
performed in correspondence of each zone, assuming the
average value for the subsequent analysis.
2.3. Residual Stress Measurement
2.3.1. The Contour Method
The contour method (CM) is a relaxation method
allowing one to evaluate the stress distribution on a
specimen section [27]. From a theoretical point of view, the
contour method is a derivation of the Bueckner’s elastic
superposition principle, which states that: “if a cracked body
subject to external loading or prescribed displacements at the
boundary has forces applied to the crack surfaces to close the
crack together, these forces must be equivalent to the stress
distribution in an uncracked body of the same geometry
subject to the same external loading” [28]. In other words
stresses to be applied on a specimen section in order to
deform the surface to its initial undeformed shape are
equivalent to the residual stresses acting on the same section
before the cut. Theoretically, all of the three stress
components acting on the cut surface can be derived,
however, evident limitations in in-plane displacements
measurement reduce the effective applicability of CM to
normal stresses [27]. The most intriguing capability of the
Fig. (2). Experimental and numerical NZ and TMAZ (Test 15). Numbers on contour lines indicated the grain size in m.
Longitudinal Residual Stress Analysis in AA2024-T3 Friction Stir Welding
method, if compared to other destructive techniques, is that a
complete residual stresses mapping can be achieved,
following a relatively simple and cheap procedure.
Moreover, the experimental steps can be realized using
devices and machines currently available in most industrial
as well as academic research laboratories. For these reasons,
CM has been widely applied in several contexts, including
FSW. Woo et al. in [29] discussed some results provided by
an experimental investigation on friction stir processing of
AZ31B magnesium alloy. ND and CM were applied for
residual stress evaluation providing similar stress
distributions. An application of CM to analyse the effect of
shot peening and laser shock peening on AA2195 and
AA7075 friction stir welds was detailed in [30], highlighting
a remarkable through the thickness variation of the
longitudinal stress. ND and CM were employed by Prime et
al. to investigate longitudinal residual stresses in AA7050T7451 and AA2024-T351 friction stir butt joints [31]. Good
agreement was found in stress mapping, evidencing also an
M-shaped longitudinal stress profile. More recently, RichterTrummer implemented the CM to analyse the effect of
clamping forces in FSW of 3.18 mm thick AA2198-T851
plates [32]. It was found that higher clamping forces lead to
more uniform residual stress distribution through the
thickness.
2.3.2. Sample Cut
The application of the contour method is based on four
consecutive steps: specimen cut, relaxed surface acquisition,
data reduction, and stress computation. Specimen sectioning
is a crucial step, since it potentially affects the shape of the
relaxed surface and, as a consequence, the reliability of the
inferred normal stress. The cutting process should respect
some prerequisites to avoid or at least reduce measurement
errors. The cut should be flat and of high surface finishing,
the cut width should be constant and very small and the
specimens should be rigidly clamped on both sides.
Additionally, the just cut (and relaxed) surfaces should not
be re-machined by the cutting tool. The cutting technology
more suitable to accomplish this step is the wire electro
discharge machining (WEDM). In the present investigation,
FSWed specimens have been sectioned in correspondence of
the mid-length and orthogonally to the weld line. The
WEDM process has been performed by means of a
MITSUBISHI FA-20 machine, using a copper 0.25 mm
diameter wire in a deionized water bath. Skim-cut
parameters have been used to minimize effects on preexisting stresses.
2.3.3. Contour Measurement
After sectioning, residual stresses are free to relax,
leading to a three dimensional displacement of the cut
surfaces. Two approaches to contour measurement are
generally applied, based, respectively, on contact (tactile) or
non-contact (optical) techniques. Optical devices are
nominally less accurate than tactile machines, such as
Coordinates Measuring Machines (CMM), however, the
improvement of the acquisition rate and the enlargement of
the dataset virtually result in accuracy comparable with
tactile systems [33].
In the present analysis, out of plane displacements of the
sectioned surfaces have been recorded by means of a CMM
The Open Mechanical Engineering Journal, 2013, Volume 7
21
(DEA IMAGE GLOBAL CLIMA), in a moisture and
temperature controlled room. The resolution of the used
CMM is 0.1 m. The machine has been equipped with a
contact probe constituted by a 30 mm height steel stylus and
a 3 mm diameter ruby tip. The machine-probe system has
been calibrated before each acquisition, to properly
compensate stylus deflection. A 0.5 mm x 0.5 mm
measurement grid has been programmed in a reference
coordinate system associated with the sample, resulting into
about 1100 data points for each surface. Taking into account
the reduced extension of the cut surface, a hen-peck
acquisition mode has been adopted to improve measurement
accuracy without excessive penalization of the acquisition
time. As aforementioned, the contour method allows for the
evaluation of normal stress only: shear stress should be
compensated by averaging the opposite displaced contours
[27]. To accomplish this prerequisite, cut surfaces on both
halves of each specimen have been measured.
2.3.4. Data Reduction
This step is relevant for the effectiveness of the whole
procedure, since stress computation magnifies measurement
errors. What is more, an analytical description of the
displaced surface is highly desirable, considering that it
allows one to define the FE mesh independently of the
measurement grid. Intuitively, nodal displacements can be
easily derived if a mathematical representation of the relaxed
surface is available. Experimental data have been imported
in the MATLAB environment and then averaged and fit to
an unique smoothing surface, significantly improving the
quality of the regression with respect to the polynomial
surface fitting adopted by the same authors in [34]. Finally,
nodal displacements from the base plane have been inferred
and exported for FE calculations. In Fig. (3a, b) the
smoothing surface approximating experimental data and the
nodal displacements are shown.
2.3.5. Stress Computation
Longitudinal residual stresses distribution has been
computed by means of an elastic FE model of the cut
sample. The digitalized out of plane displacements have
been used, with reversed sign, as input nodal boundary
conditions, assuming an initial block shaped geometry. This
procedure, commonly adopted for the application of the
contour method, is less labor intensive with respect to the
creation of the deformed model and the successive forcing of
the relaxed surface to a flat status [27-34]. Additional
constraints have been imposed on a node on the surface
opposed to the displaced section to prevent rigid body
motion. The commercial software ANSYS has been used to
solve the linear elastic boundary value problem. In Fig. (3c)
a computed deformed shape of the half specimen has been
depicted. Two stress maps have been inferred for each
processing condition. In the first case (LRS1 in what
follows) no variation of elastic properties of the processed
material is considered, assuming the parent material
properties for all the welding zones. In the second set of
simulations (LRS2), thermal and mechanical effects in the
different welding zones have been considered assigning the
ultrasonically detected Young modulus to each zone. As a
first approximation, the same set of displacements has been
used in both cases.
22 The Open Mechanical Engineering Journal, 2013, Volume 7
Carlone and Palazzo
Fig. (3). Experimental points and smoothing surface (a); base plane and nodal displacements (b); deformed shape (Test 12) due to stress
relaxation after the cut (c).
3. RESULTS AND DISCUSSION
3.1. Material Characterization and Influence
Properties Variations on Stress Computing
of
As aforementioned, during FSW material heating and
stirring induce several microstructural phenomena,
significantly affecting the final material properties. The base
material welded in the present investigation showed the
typical rolling microstructure, characterized by elongated
grains, whit mean grain size equal to 59.18 m. A
remarkable grain refinement has been found in the NZ of all
samples, ranging the recrystallized grain size between 8.43
and 11.13 min the central point, respectively in Test 3 and
Test 13 [22,24]. No significant grain size variation has been
observed in HAZ and TMAZ with respect to the parent
material. Ultrasonic measurements showed that, as relatively
to the grain size, no significant TOF (and consequently
elastic modulus) variations are induced in the HAZ and
TMAZ with respect to the base material (see Fig. 4), whose
Young modulus resulted equal to 73.3 GPa. Slight major, but
still not elevated, deviations have been found at the weld line
(in correspondence of the NZ) of all samples, being the
maximum increase equal to 3.4 GPa (4.6% with respect to
the parent material modulus) in Test 10. The positive
variation of Young modulus has been related to the grain
refinement in the NZ due to continuous dynamic
recrystallization. Differently from what already declared
relatively to microhardness [22], no evident correlation was
found between the Young modulus and process parameters.
This aspect can be related to variation of shape and extension
of the NZ, as well as the grain size distribution, with process
parameters. Considering that ultrasonically measured
properties are averaged through the thickness and that at
lower the recrystallized zone do not reach the weld root,
an increase of the TOF and a reduction of the derived elastic
modulus are expected.
In Fig. (5a-d) some results concerning the influence of
properties variations on the inferred stress field have been
depicted. In particular, Fig. (5a, b) show (in vertical
sequence) the LRS distribution computed using uniform
material properties, variable material properties, the absolute
(LRS = LRS2-LRS1) and the percentage difference
(LRS/LRS2)·100 relatively to opposite process conditions
(Test 1: minimum and ; Test 15: maximum and ). A
minimum threshold (± 2 MPa) has been adopted for the
computation of the LRS percentage difference to avoid
calculations in (practically) stress free zone, resulting in the
conditioning of the overall evaluation. In Fig. (5c) a
summary graph from all test cases, showing the maximum
LRS and the absolute maximum tensile LRS, has been
reported.
Fig. (4). Ultrasonic signal in the BM, HAZ and NZ (Test 15).
As can be seen, the LRS1 and LRS2 distributions overlap
without significant (qualitative as well as quantitative)
deviations. Indeed, the maximum stress difference (in
absolute value) ranged between about 1 and 4 MPa in all
cases, apart Test 3 (7 MPa) and Test 12 (5 Mpa). It should be
also noted that the higher differences have been found in
correspondence of relatively more stressed zone, resulting in
few percentage points of maximum deviations between
results provided by the two procedures. The above
comparison justifies the widely used assumption of uniform
material properties in residual stress evaluation following the
Longitudinal Residual Stress Analysis in AA2024-T3 Friction Stir Welding
contour method [28, 30-34], avoiding deeper metallographic
and mechanical characterization of the processed material.
However, even focusing on the residual stress field
induced by FSW process, it should be noted that the above
statement can not be generalized and extended, without
further verifications, to more specific cases, i.e. for instance
hybrid aluminum-copper, aluminum-steel, or aluminumtitanium FSW. Indeed, obtained outcomes indicate that the
maximum difference in LRS corresponds to the maximum
Young modulus increase (Test 3), implicitly suggesting the
need for a preliminary materials characterization for the
effective application of the method to hybrid welding
processes. What is more, in this case, the equilibrium
condition of force and moments acting on the cut section
could not be enforced by the equilibrium of displacements
and rotations, imposing the implementation of iterative
procedures to define the base plane position.
The LRS distribution, as provided by the contour method in
Test 6, is shown in Fig. (5d), magnifying the sample thickness
for clarity. The depicted stress field is qualitatively
representative of LRS maps computed in the whole
measurement campaign and evidences a tensile residual stress
status in correspondence of the weld line. Two different tensile
peaks have been found, localized at a distance from the weld
The Open Mechanical Engineering Journal, 2013, Volume 7
23
center approximately equal to the shoulder radius into the
advancing (AS) and retreating (RS) sides. The largest tensile
stress, equal to 145MPa, has been detected in the AS below the
top surface, in agreement with other reports [14,18,19], whereas
the tensile peak in the RS resulted equal to about 125MPa (85%
of the absolute maximum). The typical M-shaped profile has
been observed, as already detected by other researchers, testing
AA2024 and other aluminum alloys by means of several
measurement techniques [11-14,16,17,20,31,32]. The observed
asymmetry in the LRS distribution with respect to the weld line
can be easily explained considering the different relative
(processing material - tool) velocity moving from the AS to the
RS. Moreover, line plots (on the right in the same subfigure) of
LRS highlight some through the thickness stress variations in
correspondence of the tensile peak (AS). As can be seen
compressive stresses have been found also in correspondence of
the top and bottom surfaces of the welded sample, changing to
tensile at a distance from the external surface of about 0.5 mm.
A relatively less severe stress profile has been computed at the
center of the weld line. The tensile stress in the central region is
balanced by compressive residual stresses moving towards the
base material, reaching the minimum values about 20 mm away
from the weld line and then slowly approaching (in absolute
terms) lower values.
Fig. (5). Longitudinal residual stresses results, as provided by the contour method, assuming uniform material properties and variable
material properties (absolute and percentage stress difference are also shown), in correspondence of the two limit test case a) Test 1 ( = 800
rpm - = 35 mm/min, defected joint); and b) Test 15( = 1600 rpm - = 140 mm/min, sound joint). c) Bar plot of LRS results from all test
cases; d) LRS distributions and profiles at the mid-thickness and across the thickness in the advancing side in Test 6 ( = 1000 rpm - = 140
mm/min).
24 The Open Mechanical Engineering Journal, 2013, Volume 7
Carlone and Palazzo
Fig. (6). Influence of rotational speed on LRS: a) stress distribution on the transverse section; b) LRS profiles at the mid-thickness; c) LRS
profiles across the thickness in the advancing side ( = 140 mm/min in all test cases).
3.2. Influence of Process Parameters on LRS
Figs. (6, 7) graphically outline the influence of process
parameters and on LRS. Again, the reported distributions
and profiles fairly agree with results presented in the herein
cited bibliography, evidencing two peaks in the tensile
stressed zone and the presence of compression on both sides
of the weld line. Furthermore, a stress reduction is observed
at the center of the section (x = 0), preserving, in all the
examined cases, the positive sign. The absolute maximum is
always localized in the AS at a distance from the weld line
approximately equal to the shoulder radius, i.e. at higher
relative velocity and shear strain rate. The value of the
tensile peak in the RS lied in a range whose lower limit was
about the 80% of the peak value in the advancing side. In
particular, the gap between tensile peaks appears more
evident at lower , i.e. when the difference in material - tool
relative motion between the AS and the RS is higher.
Obtained outcomes, in terms of stress profiles and tensile
peaks, suggest a key role played by . As depicted in Fig.
(6b, c), a fair superimposition of stress profiles can be
observed varying in the considered range and assuming as fixed. On the other hand, the increase of induces
relatively more severe residual stress status, as highlighted
also in Fig. (7a-c). The predominant role of can be justified
keeping in mind that it directly establishes the thermal cycle
in terms of heating and cooling rates. Indeed, for a fixed
value of , similar temperature peaks are experienced by the
processing material despite consistent variations [22]. The
increase of reduces the time available for the diffusion of
the heat dissipated in the material away from the weld line.
As a consequence, higher temperature gradients are induced
and higher stresses on cooling should be expected. Inferred
distributions evidenced also the influence of the presence of
welding defects on LRS. In defected joints, at lower (Tests
1, 4, and 7), a remarkable increase of the tensile stress peak
with respect to sound joints processed using the same (Tests 10 and 13) has been detected, in agreement with other
reports [19]. This effect vanishes at higher .
The influence of on the tensile peak has been also
tested by means of simple regression analysis. In particular
linear and logarithmic regressions have been performed
considering the whole dataset (A) and reducing the dataset to
sound joints (B). A remarkable correlation coefficient R2 has
been obtained in all conditions, as summarized in Table 1.
On the other hand, a very low correlation coefficient has
been found repeating the procedure assuming as
independent variable. It should be remarked, however, that
the reduced extension of the dataset, in particular in Case B,
did not allow for a rigorous statistical analysis, therefore the
aforementioned considerations should be considered as
indicative of the key role played by . It is worth also noting
that the relatively better correlation adopting a logarithmic
regression indicates the probable approaching of a tensile
peak limit as increases, in agreement with the temperature
peak trend described in [22].
CONCLUSIONS
The longitudinal residual stress distribution in AA2024T3 friction stir welded butt joints has been investigated by
means of the contour method. Two approaches to stress
computation have been adopted, assuming, respectively,
uniform and variable material properties. Taking into
account what above reported, the following conclusions can
be highlighted:
i.
an increase of the Young modulus is induced by the
grain refinement in the NZ, however, no evident
correlation with process parameters has been
observed. Negligible elastic modulus variations have
Longitudinal Residual Stress Analysis in AA2024-T3 Friction Stir Welding
The Open Mechanical Engineering Journal, 2013, Volume 7
25
Fig. (7). Influence of welding speed on LRS: a) stress distribution on the transverse section; b) LRS profiles at the mid-thickness; c) LRS
profiles across the thickness in the advancing side ( = 1400 rpm in all test cases).
ii.
iii.
iv.
v.
been found in the HAZ and TMAZ with respect to the
parent material;
Table 1.
a detailed mapping of longitudinal residual stresses
has been achieved by means of the used procedure,
providing suitable material to validate computational
simulative models of the process and to be included
in static as well as dynamic analysis of welded
assemblies;
Dataset
the assumption of uniform material properties appears
acceptable in the investigated case, but not
generalizable. Further development of the method are
needed in order to investigate more complex cases,
such as hybrid FSW;
an asymmetric longitudinal residual stress distribution
has been found for all processing conditions,
characterized by a tensile stress in correspondence of
the weld line, balanced by compressive residual
stresses moving toward the base material. Two tensile
peaks, describing the typical M shaped stress profile,
have been measured in the tensile region, being the
minor peak (in the RS) at least the 80% of the
absolute maximum (in the AS);
the welding speed resulted the dominant process
parameter relatively to process induced longitudinal
residual stresses, since it directly affects heating and
cooling cycles. The increase of reduces the time
available for heat diffusion away from the weld line,
inducing higher temperature gradients and higher
stresses on cooling. A minor influence of the rotating
speed has been observed.
CONFLICT OF INTEREST
The authors confirm that this article content has no
conflict of interest.
A
Regression Analysis Results
Regression Type
Equation
R2
linear
0.672 + 54.64
0.90
logarithmic
52.34 ln() - 112.77
0.92
linear
0.771 + 41.93
0.93
logarithmic
61.44 ln() - 156.15
0.99
B
ACKNOWLEDGEMENTS
Declared none.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
H.N.B. Schmidt, T.L. Dickerson, and J.H. Hattel, “Material flow in
butt friction stir welds in AA2024-T3”, Acta Mater., vol. 54, pp.
1199-1209, February 2006.
M.A. Sutton, B. Yang, A.P. Reynolds, and R. Taylor,
“Microstructural studies of friction stir welds in 2024-T3
aluminum”, Mater. Sci. Eng. A Struct., vol. 323, pp. 160-166,
January 2002.
J.W. Yoon, G.H. Bray, R.A.F. Valente, and T.E.R. Childs,
“Buckling analysis for an integrally stiffened panel structure with a
friction stir weld”, Thin Wall. Struct., vol. 47, pp. 1608-1622,
December 2009.
R.M.F. Paulo, P. Carlone, R.A.F. Valente, F. Teixeira-Dias, and
G.S. Palazzo, “Integrated design and numerical simulation of
stiffened panels including friction stir welding effects”, Key. Eng.
Mater., vol. 554-557, pp. 2237-2242, April 2013
R.M.F. Paulo, F. Teixeira-Dias, and R.A.F. Valente, “Numerical
simulation of aluminium stiffened panels subjected to axial
compression: Sensitivity analyses to initial geometrical
imperfections and material properties”, Thin Wall Struct., vol. 62,
pp. 65-74, January 2013.
P.M.G.P. Moreira, A.M.P. de Jesus, A.S. Ribeiro, and P.M.S.T. de
Castro, “Fatigue crack growth in friction stir welds of 6082-T6 and
6061-T6 aluminium alloys: A comparison”, Theor. Appl. Fract.
Mech., vol. 50, pp. 81-91, October 2008.
26 The Open Mechanical Engineering Journal, 2013, Volume 7
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
P.M.G.P. Moreira, F.M.F. de Oliveira, and P.M.S.T. de Castro,
“Fatigue behaviour of notched specimens of friction stir welded
aluminium alloy 6063-T6”, J. Mater. Process. Technol., vol. 207,
pp. 283-292, October 2008.
G. Bussu and and P.E. Irving, “The role of residual stress and heat
affected zone properties on fatigue crack propagation in friction stir
welded 2024-T351 aluminium joints”, Int. J. Fatigue, vol. 25, pp.
77-88, January 2003.
R. Citarella, P. Carlone, M. Lepore, and G.S. Palazzo, “A FEMDBEM investigation of the influence of process parameters on
crack growth in aluminum friction stir welded butt joints”, Key.
Eng. Mater., vol. 554-557, pp. 2118-2126, April 2013.
A. Ali, X. An, C.A. Rodopoulos, M. W. Brown, P. O’Hara, A.
Levers, and S. Gardiner, “The effect of controlled shot peening on
the fatigue behavior of 2024-T3 aluminium friction stir welds”, Int.
J. Fatigue, vol. 29, pp.1531-1545, August 2007.
A.F. Golestaneh, Aidy Ali, and M. Zadeh, “Modelling the fatigue
crack growth in friction stir welded joint of 2024-T351 Al alloy”,
Mater. Des., vol. 30, pp. 2928-2937, September 2009.
L. Fratini, S. Pasta, and A.P. Reynolds, “Fatigue crack growth in
2024-T351 friction stir welded joints: Longitudinal residual stress
and microstructural effects”, Int. J. Fatigue, vol. 31, pp. 495-500,
March 2009.
O. Hatamleh, “A comprehensive investigation on the effects of
laser and shot peening on fatigue crack growth in friction stir
welded AA 2195 joints”, Int. J. Fatigue, vol. 31, pp. 974-988, May
2009.
M. Peel, A. Steuwer, M. Preuss, and P.J. Withers, “Microstructure,
mechanical properties and residual stresses as a function of welding
speed in aluminium AA5083 friction stir welds”, Acta Mater., vol.
51, pp. 4791-4801, September 2003.
M. Grujicic, G. Arakere, H.V. Yalavarthy, T. He, C.-F. Yen, and
B.A. Cheeseman, “Modeling of AA5083 material-microstructure
evolution during butt friction-stir welding”, J. Mater. Eng.
Perform., vol. 19, pp.672-684, July 2010.
H. Lombard, D.G. Hattingh, A. Steuwer, and M.N. James, “Effect
of process parameters on the residual stresses in AA5083-H321
friction stir welds”, Mater. Sci. Eng. A Struct., vol. 501, pp. 119124, February 2009.
K. Deplus, A. Simar, W. Van Haver, and B. de Meester, “Residual
stresses in aluminium alloy friction stir welds”, Int. J. Adv. Manuf.
Technol., vol. 56, pp. 493-504, September 2011.
L. Fratini, and B. Zuccarello, “An analysis of through-thickness
residual stresses in aluminium FSW butt joints”, Int. J. Mach. Tool.
Manuf., vol. 46, pp. 611-619, May 2006.
W. Xu, J. Liu, and H. Zhu, “Analysis of residual stresses in thick
aluminum friction stir welded butt joints”, Mater. Des., vol. 32, pp.
2000-2005, April 2011.
M.T. Milan, W.W. Bose Filho, J.R. Tarpani, A.M.S. Malafaia,
C.P.O. Silva, B.C. Pellizer, and L.E. Pereira, “Residual stress
evaluation of AA2024-T3 friction stir welded joints”, J. Mater.
Eng. Perform., vol. 16, pp. 86-92, February 2007.
M.A. Sutton, A.P. Reynolds, D.-Q. Wang, and C.R. Hubbard, “A
study of residual stresses and microstructure in 2024-T3 Aluminum
Received: April 15, 2013
Carlone and Palazzo
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
Revised: June 3, 2013
friction stir butt welds”, J. Eng. Mater.-T.ASME, vol. 124, pp. 215221, April 2002.
P. Carlone and G.S. Palazzo, “Influence of process parameters on
microstructure and mechanical properties in AA2024-T3 friction
stir welding”, Metallogr. Microstruct. Anal., DOI 10.1007/s13632013-0078-4, 2013.
C. Genevois, D. Fabrègue, A. Deschamps, and W.J. Poole, “On the
coupling between precipitation and plastic deformation in relation
with friction stir welding of AA2024 T3 aluminium alloy”, Mater.
Sci. Eng. A Struct., vol. 441, pp. 39-48, December 2006.
P. Carlone and G.S. Palazzo, “A numerical and experimental
analysis of microstructural aspects in AA2024-T3 friction stir
welding”, Key. Eng. Mater., vol. 554-557, pp. 1022-1030, April
2013.
H. Atharifar, D. Lin, and R. Kovacevic, “Numerical and
experimental investigations on the loads carried by the tool during
friction stir welding”, J. Mater. Eng. Perform., vol. 18, pp. 339350, June 2009.
R. Nandan, G.G. Roy, and T. Debroy, “Numerical Simulation of
Three-Dimensional Heat Transfer and Plastic Flow During Friction
Stir Welding”, Metall. Mater. Trans. A, vol. 37, pp. 1247-1259,
April 2006.
M.B. Prime, “Cross-sectional mapping of residual stresses by
measuring the surface contour after a cut”, J. Eng. Mater.-T.ASME,
vol. 123, pp.162-168, April 2001.
H.F Bueckner., “The propagation of cracks and the energy of
elastic deformations”, Trans. A.S.M.E., vol. 80, pp. 1225-1230,
August 1958.
W. Woo, H. Choo, M.B. Prime, Z. Feng, and B. Clausen,
“Microstructure, texture and residual stress in a friction-stirprocessed AZ31B magnesium alloy”, Acta Mater., vol. 56, pp.
1701-1711, May 2008.
O. Hatamleh and A. DeWald, “An investigation of the peening
effects on the residual stresses in friction stir welded 2195 and
7075 aluminum alloy joints”, J. Mater. Process. Technol., vol. 209,
pp. 4822-4829, June 2009.
M.B. Prime, T. Gnäupel-Herold, J.A. Baumann, R.J. Lederich,
D.M. Bowden, and R.J. Sebring, “Residual stress measurements in
a thick, dissimilar aluminum alloy friction stir weld”, Acta Mater.,
vol. 54, pp. 4013-4021, September 2006.
V. Richter-Trummer, E. Suzano, M. Beltrão, A. Roos, J.F. dos
Santos, and P.M.S.T. de Castro, Influence of the FSW clamping
force on the final distortion and residual stress field, Mater. Sci.
Eng. A Struct., vol. 538, pp. 81- 88, March 2012.
M.B. Prime, R. J. Sebring, J. M. Edwards, D. J. Hughes, and P. J.
Webster, “Laser surface-contouring and spline data-smoothing for
residual stress measurement”, Exp. Mech., vol. 44, pp.176-184,
April2004.
P. Carlone and G.S. Palazzo, “Experimental analysis of the
influence of the process parameters on residual stress in AA2024T3 friction stir welds”, Key. Eng. Mater., vol. 504-506, pp. 753758, February 2012.
Accepted: June 4, 2013
© Carlone and Palazzo; Licensee Bentham Open.
This is an open access article licensed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/)
which permits unrestricted, non-commercial use, distribution and reproduction in any medium, provided the work is properly cited.